I Will DESI confirm that the Universe isn't described by the ΛCDM model?

[Suekdccia]

TL;DR Summary

If DESI three year study shows again that dark energy is decreasing, would it be confirmed?

Tomorrow DESI new results from the 3 year study will be released (https://elements.lbl.gov/news/new-measurements-from-desi-shine-light-on-dark-energy/)

If they find again, like in their previous release of the 1st year of the study, that dark energy appears to decrease, contrary to the Lambda-CDM model, would this be officially confirmed? Or would we need more measurments to confirm whether dark energy is being reduced?

I mean, from all our measurements up to date, all indicated that dark energy is constant, so if only one study shows that it may be decreasing, even though is a very precise one, wouldn't we need more independent measurements to be sure about it?

 

[PeterDonis]

contrary to the Lambda-CDM model

A time-varying dark energy density is not "contrary" to the Lambda-CDM model in the sense of requiring the model to be scrapped. It's easy enough in the model to make the dark energy density a variable parameter. The model up to now has assumed that the dark energy is constant because (a) that's the simplest assumption, and (b) there hasn't been any evidence up to now to contradict it.

In terms of an underlying explanation for why the dark energy density varies, that would be an open area of research. But it's already an open area of research trying to find an underlying explanation for a constant dark energy density; there is no consensus answer to that. Knowing that the energy density is variable would rule out the simplest explanation (a cosmological constant that's just an inherent property of spacetime), but that's all.

 

[Orodruin]

A time-varying dark energy density is not "contrary" to the Lambda-CDM model in the sense of requiring the model to be scrapped.

That depends what you read into ”Lambda-CDM”. The Lambda would typically be interpreted as a cosmological constant and not any other form of dark energy. As such I would say that Lambda-CDM does not allow for time-varying DE density. However, you are of course correct that this is just a special case and dark energy of somewhat different equation of state is also being actively considered.

 

[Suekdccia]

That depends what you read into ”Lambda-CDM”. The Lambda would typically be interpreted as a cosmological constant and not any other form of dark energy. As such I would say that Lambda-CDM does not allow for time-varying DE density. However, you are of course correct that this is just a special case and dark energy of somewhat different equation of state is also being actively considered.

This is what I was referring to

Returning to the question then, would DESI results confirm an evolving dark energy parameter if they turned out to be right? Or would we need more measurements with other techniques to be sure?

 

[PeterDonis]

That depends what you read into ”Lambda-CDM”.

Yes, but "Lambda-CDM" is just a name. It's easy enough to change the name if the dark energy density turns out not to be constant. The model itself can accommodate that, so while a variable dark energy density might technically be "contrary" to the "Lambda" part of the name, it's not "contrary" to the model in any significant sense. It's just an adjustment to the model that would have to be made in view of more data coming in; that's the sort of adjustment that scientists make to models all the time.

 

[PeterDonis]

would DESI results confirm an evolving dark energy parameter if they turned out to be right?

Isn't that what "turned out to be right" means, by definition?

Or would we need more measurements with other techniques to be sure?

Doesn't "turned out to be right" rule that out, by definition?

You seem to have some kind of idea that there's a light that goes on that tells us for sure whether a result is "right", or whether we need to go out and do more checking. That's not how it works.

 

[Suekdccia]

Isn't that what "turned out to be right" means, by definition?


Doesn't "turned out to be right" rule that out, by definition?

You seem to have some kind of idea that there's a light that goes on that tells us for sure whether a result is "right", or whether we need to go out and do more checking. That's not how it works.

No

I'm clearly not expressing myself correctly today. By "turned out to be right" I meant to say that their fidnings turned out to be statistically sifgnificant.

What I'm referring to is that in science we usually need multiple measurements from independent sources to confirm a finding, we don't just rely on one source.

My question is actually simple: If DESI finds on its three year data that a decreasing dark energy parameter is statistically significant, could we rely on it due to its precision? Or we would need other independent measurements to confirm it as it is usually done in science?

 

[PeterDonis]

My question is actually simple

Yes, and it's still basically the same thing I said before: you think that there's a definite yes/no answer to your question. There isn't. That's not how it works. The only answer is "we'll see".

A better question might be: will cosmologists start modifying their models of the universe if the DESI results turn out to be statistically significant? But even that probably doesn't have a single definite yes/no answer; probably some cosmologists will start working on revised models, while others will opt to wait for more information.

 

[Suekdccia]

Yes, and it's still basically the same thing I said before: you think that there's a definite yes/no answer to your question. There isn't. That's not how it works. The only answer is "we'll see".

A better question might be: will cosmologists start modifying their models of the universe if the DESI results turn out to be statistically significant? But even that probably doesn't have a single definite yes/no answer; probably some cosmologists will start working on revised models, while others will opt to wait for more information.

Fair enough

 

[stephaneww]

Dark energy, when theorised by the cosmological constant, is and will remain a constant. Albert's theory is not so easily disproved:

Using the prediction of the Hubble constant value H, demonstrated by Espen Haug, on the basis of the CMB temperature (2.72548 K) and Rh=c/H, we find H = 2.1679 10-18 s-1
https://link.springer.com/article/10.1007/s10773-024-05570-6
https://hal.science/hal-04269991v1/document

T_H0=Tcmb

for Λ=1.1056 10-52 m-2, we have
ΩΛ= Λ c2 / ( 3 H2 )

ΩΛ = 0.70476, hence
Ωm = 0.29524

The DESI results give as a constraint to the ΛCDM model (cf equ17, https://arxiv.org/pdf/2503.14738 )

Ωm = 0.295

I don't see what the problem is. Albert is always right.

Stéphane Wojnow

 

[stephaneww]

Well, I'm afraid I have to do my mea culpa : the results of the DESI instrument are irrevocable: in the ΛCDM model, this variation in the cosmological constant must be taken into account. My mistake was to limit myself to its value today, which is consistent with both the ΛCDM model and the proposed Rh=ct model.

 

[ohwilleke]

would DESI results confirm an evolving dark energy parameter if they turned out to be right? Or would we need more measurements with other techniques to be sure?

We would need more measurements, and also a theory to explain the results.

In physics, the usual threshold for declaring that a new scientific discovery has been made is an experiment or observation with at least a five standard deviation statistical significance (commonly called five sigma).

The very high five sigma threshold that is used in physics as a cutoff for a "discovery" is a crude solution to deal with the fact that observational and experimental uncertainties that are treated for convenience as if they have a "normal" a.k.a. "Gaussian" distribution, actually have been shown empirically to have fatter tails than that making high sigma events more likely, to deal with complicated to quantify "look elsewhere effects" that make improbable events likely to happen somewhere if you make enough observations, and to deal with the fact that systemic error is easy to underestimate. Sophisticated and seasoned physicists don't really believe that five sigma events are as profoundly improbable as a naive mathematical calculation using a normal probability distribution would imply.

But five sigma significance isn't the only requirement for a discovery of new physics to be widely accepted.

One cannot be sure about any major scientific conclusion, no matter how great its statistical significance (and the DESI results are mostly sub-five sigma, or just barely over five sigma with possible disputes over how to calculate that properly), until (1) it is credibly replicated independently (and not credibly contradicted), and (2) someone can propose some sort of internally consistent scientific theory that can explain the new data.

Until then, there are only stronger and weaker anomalies in the observational data.

Discovering and replicating an anomaly at a five sigma level is a very big deal. But even then, until you can come up with some reasonable theoretical explanation for the anomaly with new physics, you still aren't there and will struggle to secure widespread acceptance for the conclusion that the leading legacy theory is wrong.

Unexplained large anomalies that have no reasonable theoretical explanation, realistically, often turn out to be shared systemic errors at the end of the day.

Of course, the standard I've laid out is just a social consensus among physicists about what is good enough for the community to agree upon that is informed by their experiences as a discipline. There is no magic threshold in Nature that says this degree of statistical certainty is certain, and this degree of statistical certainty is not good enough. Intrinsically, every scientific observation comes with an uncertainty attached to it and a domain of applicability, no matter how widely accepted it is in the physics community. Scientists have collectively just drawn some arbitrarily and socially constructed lines regarding what they think is adequate to meet the burden of scientific proof for their purposes, because it makes their lives easier and makes achieving consensus easier, as a practical matter.

 

[Orodruin]

to deal with complicated to quantify "look elsewhere effects" that make improbable events likely to happen somewhere if you make enough observations

How is the look-elsewhere effect complicated to quantify?

 

[ohwilleke]

How is the look-elsewhere effect complicated to quantify?

There are simple cases, where you have comparable searches for something done X times.

But it isn't always obvious how to determine how many "looks" there have been, particularly when the searches that are done have some overlap with each other but aren't identical.

 

[mitchell porter]

One reason that constant dark energy has a grip on people's imaginations, is the idea that dark energy is vacuum energy. Conventional theory suggests that there should be a constant vacuum energy (from the ground state of all the quantum fields) that gravitates, so it's easy to assume that that's what dark energy is.

 

[Orodruin]

Conventional theory suggests that there should be a constant vacuum energy (from the ground state of all the quantum fields) that gravitates, so it's easy to assume that that's what dark energy is.

It also gets the value wrong by ”a couple” of orders of magnitude …

 

[javisot]

Saying that dark energy is vacuum energy is problematic, considering that when calculating the value of that vacuum energy with our best theorys we arrive to the worst prediction in the history of physics, "a couple" of orders of magnitude as Orodruin sarcastically says. https://en.wikipedia.org/wiki/Cosmological_constant_problem

 

[Jorrie]

Well, I'm afraid I have to do my mea culpa : the results of the DESI instrument are irrevocable: in the ΛCDM model, this variation in the cosmological constant must be taken into account. My mistake was to limit myself to its value today, which is consistent with both the ΛCDM model and the proposed Rh=ct model.

From the latest DESI data release papers, the w₀w_aCDM model fits the data best. I found it enlightening to look at the H(a) equation options like this:

The complicated looking w₀w_a exponents are discussed in https://arxiv.org/pdf/2503.14738 page 5.

 

[ohwilleke]

One reason that constant dark energy has a grip on people's imaginations, is the idea that dark energy is vacuum energy. Conventional theory suggests that there should be a constant vacuum energy (from the ground state of all the quantum fields) that gravitates, so it's easy to assume that that's what dark energy is.

Another reason is that a cosmological constant is a minimal modification of Einstein's Field Equations of General Relativity that can be made seamlessly without disrupting the mathematical integrity of the EFE. This has been considered a plausible and simple variant of GR without a cosmological constant for more than a century (a decade before the astronomy observations supporting some form of dark energy started to be made).

While the term "dark energy" coined in 1998, is suggestive of some sort of substance (as a counterpoint to the almost sixty years older term "dark matter"), it is in essence a gravitational modification, and not a substance, in the form used in LambdaCDM.

It is also attractive because alternatives inject more observationally determined physical constants into the cosmology model than a single cosmological constant, which is, at a minimum, a decent first order approximation of what is observed. One of the strong points of LambdaCDM as a model is that it comes reasonably close to explaining so much of the history of the universe with so few free parameters (just six, with the cosmological constant calculated from those observables, in its most bare bones version). Additional dark energy parameters detract from that, although they may be necessary because Nature doesn't always act the way that we want it to.